Floor M van Oudenhoven1,2, Sophie H N Swinkels3, Hilkka Soininen4,5, Miia Kivipelto4,6,7,8,9, Tobias Hartmann10,11, Dimitris Rizopoulos12. 1. Department of Biostatistics, Erasmus Medical Center, PO Box 2040, 3000, Rotterdam, CA, the Netherlands. floor.van-oudenhoven@danone.com. 2. Danone Nutricia Research, Uppsalalaan 12, 3584 CT, Utrecht, The Netherlands. floor.van-oudenhoven@danone.com. 3. Danone Nutricia Research, Uppsalalaan 12, 3584 CT, Utrecht, The Netherlands. 4. Department of Neurology, Institute of Clinical Medicine, University of Eastern Finland, PO Box 1627, 70211, Kuopio, Finland. 5. Neurocenter, Department of Neurology, Kuopio University Hospital, PO Box 100, 70029, Kuopio, Finland. 6. Division of Clinical Geriatrics, Department of Neurobiology, Care Sciences and Society, Karolinska Institute, 14157, Huddinge, Sweden. 7. Clinical Trials Unit, Theme Aging, Karolinska University Hospital, 14152, Huddinge, Sweden. 8. Institute of Public Health and Clinical Nutrition, University of Eastern Finland, P.O. Box 1627, 70211, Kuopio, Finland. 9. Ageing Epidemiology Research Unit, School of Public Health, Imperial College London, St Dunstan's Road, London, UK. 10. Deutsches Institut für Demenz Prävention (DIDP), Medical Faculty, Saarland University, Kirrbergerstraße, 66421, Homburg, Germany. 11. Department of Experimental Neurology, Saarland University, Kirrbergerstraße, 66421, Homburg, Germany. 12. Department of Biostatistics, Erasmus Medical Center, PO Box 2040, 3000, Rotterdam, CA, the Netherlands.
Abstract
BACKGROUND: Missing data can complicate the interpretability of a clinical trial, especially if the proportion is substantial and if there are different, potentially outcome-dependent causes. METHODS: We aimed to obtain unbiased estimates, in the presence of a high level of missing data, for the intervention effects in a prodromal Alzheimer's disease trial: the LipiDiDiet study. We used a competing risk joint model that can simultaneously model each patient's longitudinal outcome trajectory in combination with the timing and type of missingness. RESULTS: Using the competing risk joint model, we were able to provide unbiased estimates of the intervention effects in the presence of the different types of missingness. For the LipiDiDiet study, the intervention effects remained statistically significant after this correction for the timing and type of missingness. CONCLUSION: Missing data is a common problem in (Alzheimer) clinical trials. It is important to realize that statistical techniques make specific assumptions about the missing data mechanisms. When there are different missing data sources, a competing risk joint model is a powerful method because it can explicitly model the association between the longitudinal data and each type of missingness. TRIAL REGISTRATION: Dutch Trial Register, NTR1705 . Registered on 9 March 2009.
BACKGROUND: Missing data can complicate the interpretability of a clinical trial, especially if the proportion is substantial and if there are different, potentially outcome-dependent causes. METHODS: We aimed to obtain unbiased estimates, in the presence of a high level of missing data, for the intervention effects in a prodromal Alzheimer's disease trial: the LipiDiDiet study. We used a competing risk joint model that can simultaneously model each patient's longitudinal outcome trajectory in combination with the timing and type of missingness. RESULTS: Using the competing risk joint model, we were able to provide unbiased estimates of the intervention effects in the presence of the different types of missingness. For the LipiDiDiet study, the intervention effects remained statistically significant after this correction for the timing and type of missingness. CONCLUSION: Missing data is a common problem in (Alzheimer) clinical trials. It is important to realize that statistical techniques make specific assumptions about the missing data mechanisms. When there are different missing data sources, a competing risk joint model is a powerful method because it can explicitly model the association between the longitudinal data and each type of missingness. TRIAL REGISTRATION: Dutch Trial Register, NTR1705 . Registered on 9 March 2009.
Authors: Martijn C de Wilde; Botond Penke; Eline M van der Beek; Almar A M Kuipers; Patrick J Kamphuis; Laus M Broersen Journal: J Alzheimers Dis Date: 2011 Impact factor: 4.472
Authors: P S Aisen; S Andrieu; C Sampaio; M Carrillo; Z S Khachaturian; B Dubois; H H Feldman; R C Petersen; E Siemers; R S Doody; S B Hendrix; M Grundman; L S Schneider; R J Schindler; E Salmon; W Z Potter; R G Thomas; D Salmon; M Donohue; M M Bednar; J Touchon; B Vellas Journal: Neurology Date: 2010-12-22 Impact factor: 9.910
Authors: Philip Scheltens; Jos W R Twisk; Rafael Blesa; Elio Scarpini; Christine A F von Arnim; Anke Bongers; John Harrison; Sophie H N Swinkels; Cornelis J Stam; Hanneke de Waal; Richard J Wurtman; Rico L Wieggers; Bruno Vellas; Patrick J G H Kamphuis Journal: J Alzheimers Dis Date: 2012 Impact factor: 4.472
Authors: Diane Jansen; Valerio Zerbi; Ilse A C Arnoldussen; Maximilian Wiesmann; Anne Rijpma; Xiaotian T Fang; Pieter J Dederen; Martina P C Mutsaers; Laus M Broersen; Dieter Lütjohann; Malgorzata Miller; Leo A B Joosten; Arend Heerschap; Amanda J Kiliaan Journal: PLoS One Date: 2013-09-24 Impact factor: 3.240
Authors: Raj C Shah; Patrick J Kamphuis; Sue Leurgans; Sophie H Swinkels; Carl H Sadowsky; Anke Bongers; Stephen A Rappaport; Joseph F Quinn; Rico L Wieggers; Philip Scheltens; David A Bennett Journal: Alzheimers Res Ther Date: 2013-11-26 Impact factor: 6.982